Prove some identities! (Round two)

**Drexel28** · December 30th 2010, 11:16 AM

And you thought it was over?

Problem: Compute

$\displaystyle \sum_{n=2}^{\infty}\left(\zeta(n)-1\right)$

,

$\displaystyle \sum_{n=1}^{\infty}\left(\zeta(2n)-1\right)$

and,

$\displaystyle \sum_{n=2}^{\infty}\left(\zeta(2n-1)-1\right)$

**Unbeatable0** · December 30th 2010, 12:33 PM

Cool ones!

$\displaystyle{ \sum_{n=2}^\infty (\zeta(n)-1) = \sum_{n=2}^\infty\sum_{m=2}^\infty \frac{1}{m^n} = \sum_{m=2}^\infty\sum_{n=2}^\infty \frac{1}{m^n} }$

$\displaystyle{= \sum_{m=2}^\infty \frac{1}{m^2-m} = \sum_{m=2}^\infty \left(\frac{1}{m-1}-\frac{1}{m}\right) = 1 }$

Likewise:

$\displaystyle{ \sum_{n=1}^\infty (\zeta(2n)-1) = \sum_{m=2}^\infty\sum_{n=1}^\infty \frac{1}{m^{2n}} = \sum_{m=2}^\infty\frac{1}{m^2-1} }$

$\displaystyle{ = \frac{1}{2}\sum_{m=2}^\infty \left(\frac{1}{m-1}-\frac{1}{m+1}\right) = \frac{3}{4} }$

And the last one is the difference between the first and the second, that is:

$\displaystyle{ \sum_{n=1}^\infty (\zeta(2n+1)-1) = \frac{1}{4} }$

**Drexel28** · December 30th 2010, 12:42 PM

Originally Posted by Unbeatable0

Cool ones!

$\displaystyle{ \sum_{n=2}^\infty (\zeta(n)-1) = \sum_{n=2}^\infty\sum_{m=2}^\infty \frac{1}{m^n} = \sum_{m=2}^\infty\sum_{n=2}^\infty \frac{1}{m^n} }$

$\displaystyle{= \sum_{m=2}^\infty \frac{1}{m^2-m} = \sum_{m=2}^\infty \left(\frac{1}{m-1}-\frac{1}{m}\right) = 1 }$

Likewise:

$\displaystyle{ \sum_{n=1}^\infty (\zeta(2n)-1) = \sum_{m=2}^\infty\sum_{n=1}^\infty \frac{1}{m^{2n}} = \sum_{m=2}^\infty\frac{1}{m^2-1} }$

$\displaystyle{ = \frac{1}{2}\sum_{m=2}^\infty \left(\frac{1}{m-1}-\frac{1}{m+1}\right) = \frac{3}{4} }$

And the last one is the different between the first and the second, that is:

$\displaystyle{ \sum_{n=1}^\infty (\zeta(2n+1)-1) = \frac{1}{4} }$

Right, not too hard, but interesting indeed. Of course using the same methodology shows that $\displaystyle \sum_{n=1}^{\infty}(-1)^n\left(\zeta(n)-1\right) =\frac{1}{2}$ . This one is slightly harder. Try

$\displaystyle \sum_{n=2}^{\infty}\frac{\zeta(n)-1}{n}$

**Unbeatable0** · December 30th 2010, 12:53 PM

Originally Posted by Drexel28

Right, not too hard, but interesting indeed. Of course using the same methodology shows that $\displaystyle \sum_{n=1}^{\infty}(-1)^n\left(\zeta(n)-1\right) =\frac{1}{2}$ . This one is slightly harder. Try

$\displaystyle \sum_{n=2}^{\infty}\frac{\zeta(n)-1}{n}$

Of course, the anternating one is just the difference between the second and the last ones you proposed in the first post.

I find it very interesting that they're rational.

The other one you proposed is the same method:

$\displaystyle{ \sum_{n=2}^{\infty}\frac{\zeta(n)-1}{n} = \sum_{m=2}^\infty\sum_{n=2}^\infty \frac{1}{nm^n} = \sum_{m=2}^\infty\left(\ln m - \ln (m-1) - \frac{1}{m}\right) }$

$\displaystyle{ =1+\lim_{n\rightarrow\infty} \left(\ln(n)-\sum_{m=1}^n\frac{1}{m}\right) = 1-\gamma }$

**Drexel28** · December 30th 2010, 12:55 PM

Originally Posted by Unbeatable0

Of course, the anternating one is just the difference between the second and the last ones you proposed in the first post.

I find it very interesting that they're rational.

The other one you proposed is the same method:

$\displaystyle{ \sum_{n=2}^{\infty}\frac{\zeta(n)-1}{n} = \sum_{m=2}^\infty\sum_{n=2}^\infty \frac{1}{nm^n} = \sum_{m=2}^\infty\left(\ln m - \ln (m-1) - \frac{1}{m}\right) }$

$\displaystyle{ =1+\lim_{n\rightarrow\infty} \left(\ln(n)-\sum_{m=1}^n\frac{1}{m}\right) = 1-\gamma }$

Right. How about this one

$\displaystyle \sum_{n=1}^{\infty}\frac{\zeta(2n)}{n(2n+1)}$

**chiph588@** · December 30th 2010, 02:11 PM

Originally Posted by Drexel28

Right. How about this one

$\displaystyle \sum_{n=1}^{\infty}\frac{\zeta(2n)}{n(2n+1)}$

Spoiler:

**Drexel28** · December 30th 2010, 02:33 PM

Originally Posted by chiph588@

We know $\displaystyle \frac{\sin\pi x}{\pi x} = \prod_{n=1}^\infty \left(1-\frac{x^2}{n^2}\right) \implies \log(\pi x) - \log(\sin\pi x) = \sum_{n=1}^\infty \frac{\zeta(2n)}{n}x^{2n}$ .

Thus we see $\displaystyle \sum_{n=1}^{\infty}\frac{\zeta(2n)}{n(2n+1)} = \int_0^1 \log(\pi x) dx - \int_0^1 \log(\sin\pi x) dx = \log\pi-1+\log2 = \log(2\pi)-1$ .

Right! Now someone else post a problem so I can solve one! haha

**Unbeatable0** · December 30th 2010, 02:40 PM

Here are some nice ones. Compute:

$\displaystyle \sum_{n=1}^\infty \frac{|\mu (n)|}{n^s}$

$\displaystyle \sum_{n=1}^\infty \frac{\sigma(n)}{n^s}$

$\displaystyle \sum_{n=1}^\infty \frac{2^{\omega(n)}}{n^s}$

where $\mu(n)$ is the Mobius function, $\sigma(n)$ is the divisor function, and $\omega(n)$ counts the number of distinct primes dividing $n$ .

Edit: just to clarify, by $\sigma(n)$ I was referring, as usually denoted by this letter, to the function returning the sum of the divisors of $n$ , and not the number of divisors, which is also called the divisor function.

**chiph588@** · December 30th 2010, 03:00 PM

Originally Posted by Unbeatable0

$\displaystyle \sum_{n=1}^\infty \frac{2^{\omega(n)}}{n^s}$

Note that we require $\text{Re}(s) > 1$ . I will do one, because they all can be derived in a similar fashion.

Spoiler:

Try this one: Compute

$\displaystyle \int_0^1 (-\log x)^a dx$

**Random Variable** · December 30th 2010, 03:08 PM

What's required to switch the order of summation? Absolute convergence of the inner sum?

Originally Posted by Unbeatable0

Cool ones!

$\displaystyle{ \sum_{n=2}^\infty (\zeta(n)-1) = \sum_{n=2}^\infty\sum_{m=2}^\infty \frac{1}{m^n} = \sum_{m=2}^\infty\sum_{n=2}^\infty \frac{1}{m^n} }$

$\displaystyle{= \sum_{m=2}^\infty \frac{1}{m^2-m} = \sum_{m=2}^\infty \left(\frac{1}{m-1}-\frac{1}{m}\right) = 1 }$

Likewise:

$\displaystyle{ \sum_{n=1}^\infty (\zeta(2n)-1) = \sum_{m=2}^\infty\sum_{n=1}^\infty \frac{1}{m^{2n}} = \sum_{m=2}^\infty\frac{1}{m^2-1} }$

$\displaystyle{ = \frac{1}{2}\sum_{m=2}^\infty \left(\frac{1}{m-1}-\frac{1}{m+1}\right) = \frac{3}{4} }$

And the last one is the difference between the first and the second, that is:

$\displaystyle{ \sum_{n=1}^\infty (\zeta(2n+1)-1) = \frac{1}{4} }$

**simplependulum** · December 30th 2010, 05:51 PM

$\displaystyle \sum_{n=1}^{\infty} \frac{\sigma(n) }{n^s }$

I think all of us would factorise the summation as this arithmetic function is multiplicative !

$\displaystyle = \prod_{\text{all primes p }} \left( 1 + \frac{\sigma(p)}{p^s} + \frac{\sigma(p^2)}{p^{2s}} + ... \right)$

$\displaystyle = \prod_{\text{all primes p }} \left[ \sum_{k=0}^{\infty} \left(\frac{ p^{k+1} - 1 }{p^{ks}(p - 1) } \right) \right]$

$\displaystyle = \prod_{\text{all primes p }} \frac{1}{p-1} \left( \frac{p}{1 - 1/p^{s-1} } - \frac{1}{1 - 1/p^s } \right)$

$\displaystyle = \prod_{\text{all primes p }} \frac{1}{p-1} \left( \frac{p-1}{ (1 - 1/p^{s-1} )(1 - 1/p^s) } \right)$

$\displaystyle = \prod_{\text{all primes p }} \frac{1}{(1 - 1/p^{s-1} )(1 - 1/p^s) }$

$\displaystyle = \zeta(s) \zeta(s-1)$

About the first problem , I remember it was first discovered by Ramanujan , he expanded the product :

$\displaystyle \frac{1}{\zeta(s)} = \prod_{\text{all primes p }} \left( 1 - \frac{1}{p^s} \right)$

As each factor provides only a prime but not a power of it , all the summands of the expansion should be square-free , and the signs of them exactly indicate the moebius function .

Therefore , $\displaystyle \frac{1}{\zeta(s) } = \sum_{n=1}^{\infty} \frac{\mu(n)}{n^s}$

But this time , absolute value is taken :

$\displaystyle \sum_{n=1}^{\infty} \frac{|\mu (n)|}{n^s} = \prod_{\text{all primes p }} \left( 1 + \frac{1}{p^s} \right) = \frac{ \zeta(s)}{\zeta(2s)}$

**Random Variable** · December 30th 2010, 07:22 PM

$\int^{1}_{0} (-\ln x)^{a} \ dx$

$= (-1)^{a} \int^{1}_{0} (\ln x)^{a} \ dx$

let $x = e^{-u}$

$= (-1)^{a} \int_{\infty}^{0} (-u)^{a} (-e^{-u}) \ du$

$= \int^{\infty}_{0} u^{a} e^{-u} \ du$

$= \Gamma (a+1)$

**chiph588@** · December 30th 2010, 08:59 PM

Evaluate $\displaystyle \int_0^1 \int_0^1 \frac{\log(xy)}{xy-1} dxdy$ .

**Drexel28** · December 30th 2010, 09:26 PM

Originally Posted by chiph588@

Evaluate $\displaystyle \int_0^1 \int_0^1 \frac{\log(xy)}{xy-1} dxdy$ .

Damnit, I had to go and all the cool number theory ones were already done!

Spoiler:

**simplependulum** · December 30th 2010, 10:50 PM

Spoiler:

How about these ?

$\displaystyle \int_0^1 \int_0^1 \int_0^1 \frac{dxdydz}{(1 + x^2 + y^2 + z^2)^2 }$ and

$\displaystyle M(x,y) = \frac{\pi}{4} ~ \frac{x+y}{ \big K\left[ \left( \frac{x-y}{x+y} \right)^2 \right] }$

where $\displaystyle K(m) = \int_0^{\pi/2} \frac{dx}{\sqrt{ 1 - m \sin^2(x) } }$ and $M(x,y)$ is arithmetic geometric mean :

$\displaystyle a_1 = \frac{x+y}{2} ~,~ g_1 = \sqrt{xy}$ Then

$\displaystyle a_{n+1} = \frac{a_n + g_n }{2} ~,~ g_{n+1} = \sqrt{a_n g_n }$

$\displaystyle M(x,y) = \lim_{n\to \infty} a_n = \lim_{n\to \infty} g_n$

EDIT : I have made a very big mistake on the first problem , the integrand should be squared !

Thread: Prove some identities! (Round two)

LinkBack

Thread Tools

Display

Prove some identities! (Round two)

Similar Math Help Forum Discussions

Can't prove these identities?

prove the following identities.

Prove Identities

Prove these identities.

help Prove identities

Search Tags